Blackmagic Camera Update July 2020

[Les Hillis]

Looks nice. Wish I'd turned off the mics now :-)

JB

Thanks. I should have cut the audio, but I didn't listen to it until after I uploaded it, and didn't want to use the data re-uploading.


That sunset sequence looks really good.


I get the impression from what I've seen so far that Blackmagic have reached a similar stage of development as RED did when they released the Dragon sensor. While there's more to user-adoption than just image-quality alone, it seems to me the Ursa Mini Pro 12K is a viable digital-cinema-camera alternative.

I think it's good to see another 'stock' or 'flavour' out there for people to be able to choose from and make use of.

[rand thompson]

I'm finding working with DCTLs instead of a Lut made by Resolve to give better results for Gen 5.

[rand thompson]

Here's a brief test of the scene above I removed a little more yellow from the skin in the DCTL, it's untitled. The free Vimeo account is limited to 500mb per week so I had to keep the file down. let me know if this is visible or not, haven't used vimeo in a while.

you can download it at the vimeo web page
https://vimeo.com/455686151

4K DCI

[Paul Jonathan]

Here's some more footage. It's possibly a bit dry, but a sequence of shots in the order they were shot of a sunset.

The show I'm prepping for needs a lot of weather changing and time passing footage. I shot this over two hours in 12K Q3 Blackmagic RAW. Was about 227Gb and I cut this in about 3 hours.

There's NO GRADE here, just applied "extended video". There's more DR in the BRAW files, but it was an interesting exercise in just shooting and editing to a fast turnaround.

I shot well after sunset (about 2 mins in) and at 4 mins the colours get very interesting.

https://vimeo.com/455155949

JB

Very interesting test. Great to have so much movement and fine color gradients in the same frame. The brightness jumps, are those due to different speeds of the lenses used or just because of the movement of the clouds? Because in the description you said these were edited chronologically, right? Also while I am very impressed overall, the wide shots seems to pop a little more than the long ones. What focal lengths did you use? Could this possibly be due to atmosphere haze? Did you use any UV filters?

Thanks for all the material and insight you are providing! This looks like a hell of an exciting camera.

[John Brawley]

Very interesting test. Great to have so much movement and fine color gradients in the same frame. The brightness jumps, are those due to different speeds of the lenses used or just because of the movement of the clouds? Because in the description you said these were edited chronologically, right? Also while I am very impressed overall, the wide shots seems to pop a little more than the long ones. What focal lengths did you use? Could this possibly be due to atmosphere haze? Did you use any UV filters?

Thanks for all the material and insight you are providing! This looks like a hell of an exciting camera.

Hi.

Obviously over two hours the exposure changes a lot. I was chasing it, started with LEE ND2.1, then switched to 1.2 and then nothing.

It wasn’t corrected other than the SOC look that was applied so you’re seeing dailies.

I used CP3 and some CP2 lenses.

The wides below 25mm are CP3. I had 15mm, 18mm and 21mm and then a lot of the long lenses are 135mm and 100mm. A couple of 85 and 50mm shots too.

The weather and islands are quite far away, something like 3-4 miles. It’s hard to pixel peep sharpness and resolving power on a cloud :-)

JB

[Les Hillis]

Here's a brief test of the scene above I removed a little more yellow from the skin in the DCTL...let me know if this is visible or not...

Looks really clean and colourful Rand, reminds me of something you'd see in an ad.

Had another look into this to see if I could show some less contrasty skin tones and address some of the noise I was seeing in the shadows -



The image seems flexible enough, it's interesting trying to figure out how to work with it.

[rand thompson]

Les,

That's actually what Influences all my grades, fashion magazine ads as well as the look of some movies of Italian Cinema. I like this image of yours better. But it was more appropriate to crush the shadow area by his Right ear like in your grade but maybe not as much as in your original. If you look at the right ear of my grade of this gentleman , you'll see some noise in his right ear and the right side of his neck. I believe that this was from using some parameters of the DCTL that might have been in conflict with each other in this grade. I'm trying to track it down.

The biggest problem I have been trying to deal with in both the Red Komodo and the Ursa Mini Pro 12K is a magenta cast/dominance. Maybe both will be rectified with a firmware update or better Transform Lut.

[Les Hillis]

Not sure if you're talking about the same thing, but I always fiddle with the Tint control anyway (even if I just end up leaving it at 0), so an overall magenta/green bias doesn't make much difference to me. Bonus points if the shot has a white or mid-grey reference in it that I can auto-white-balance off to get the tint value (while discarding the other values it gives).

Still not sure about some other things I'm seeing with this BRAW Ursa 12K imagery either. Nothing bad, but it's clearly doing it's own thing compared to IPP2 R3D's. Hopefully some of the post experts will dig into it and share their findings.

[rand thompson]

Tried to get close to recreating the look of the DCTL script with Resolve's Primaries and BM Gen 5 Raw settings.

[rand thompson]

Not sure if you're talking about the same thing, but I always fiddle with the Tint control anyway (even if I just end up leaving it at 0), so an overall magenta/green bias doesn't make much difference to me. Bonus points if the shot has a white or mid-grey reference in it that I can auto-white-balance off to get the tint value (while discarding the other values it gives).

Still not sure about some other things I'm seeing with this BRAW Ursa 12K imagery either. Nothing bad, but it's clearly doing it's own thing compared to IPP2 R3D's. Hopefully some of the post experts will dig into it and share their findings.

Les,

I also almost never use tint to correct the image because of the either Magenta or Green bias it leaves in the Neutral colors. I never like to get use to using Auto white balancing. There will be images that won't have either a good white, grey or black point. So it's best to get into the practice of white balancing the images manually.

I was thinking that maybe since the sensor has an equal number of Red, Green and Blue pixels, that now both the Red and Blue pixels outnumber the Green Pixels 2:1 where as with most sensors the Green pixels are dominate in relation to Red and Blue. And when Red and Blue are Dominant to Green in either the Highlights, Midtones and Shadows, Magenta is the By-product.